This proposal is to support a collaboration between the laboratories of John Allman and Barbara Wold to investigate gene expression in laser micro-dissected cell populations in autopsy brains of well-phenotyped autistic individuals versus age and sex matched controls using a recently developed techniques, RNA-Seq. RNA-Seq will be interpreted in the context of SNP and copy number variation (CNV) genotyping. FI (Fronto Insular Cortex) and ACC (Anterior Cingulate cortex) are functionally implicated in social interaction and reciprocity, in empathy, and in the awareness and regulation of bodily functions. These functions are crucially affected in autism. Unlike other homeostatic systems, including those in the hypothalamus, the FI/ACC system appears to have direct access to consciousness and motivation. Analysis of initial RNA-Seq on FI samples produced two coherent gene networks that differ between autistics and neurotypical controls. Immunostaining for several network members has begun to show how networks map onto the cellular circuitry. Specifically, at the cell level, FI and ACC contain large bipolar cells (Von Economo neurons, VENs) that are distinctive features of these structures in apes and humans. We found that VENs express receptors for the cytokines interleukin 4 (IL4R) and interleukin 6 (IL6R) in normal and autistic subjects, and RNA-Seq identified a prominent gene network related to inflammation which is strongly up-regulated in a subset of our autistic cases (autism-A). This network is centers on IL6 and includes ATF3, SOCS3, and GADD45B, which are selectively expressed in VENs. Microglia, the immune cells of the nervous system, are numerous and are in the activated state in autism-A, making a likely signal source. Our remaining autistic cases comprise an autism-B group, which is characterized by increased expression of genes in the presynaptic terminal including NRXN1 (neurexin 1), which provides Velcro-like binding to neuroligins in the post-synaptic membrane. NRXN1 is one of the genes most strongly and consistently associated with autism. NRXN1 has many splice variants which could provide specificity in formation or strength of synaptic connection. To probe these networks more deeply;to assign gene expression and splice isoforms to their proper cells;and to discover remaining differences in autistics, we propose generation-2 RNA-seq on laser-dissected cells. Thus VENs and other key cell types are rare (<5% of cells in FI). This reduces their transcriptome completeness: genes under- expressed in autism relative to controls will be especially affected. Successful laser capture can overcomes this hurdle. We also propose deep RNA sequencing for NRXN1 isoforms and other complex families that may be uncovered. PUBLIC HEALTH RELEVANCE: We seek to understand the cellular bases of autism by using a new technology, RNA-Seq, to determine differences in gene expression in autopsy brains of subjects with well described autism versus age and sex matched neurotypical individuals. We have investigated two specific cortical areas involved in self-awareness and social reciprocity which are abnormal in autism and have found increased expression in a network of genes related to inflammation in autism group A, whereas the remaining cases, autism group B, have increased expression in a network of genes related to synapses. We propose to use laser micro-dissection to investigate gene expression in specific neuronal and non-neuronal populations in the cortical areas of interest in the autism-A, autism-B and control groups.